The long-term goal of Project 2 is to define the genomic and transcriptional alterations that contribute to acute myeloid leukemia (AML) relapse after chemotherapy and/or allogeneic hematopoietic stem cell transplantation (alloHSCT). Although standard induction chemotherapy (cytarabine with an anthracycline) is curative for some AML patients, approximately 60-80% will either relapse after achieving a first complete remission, or have disease that is refractory to initial induction chemotherapy. The curative potential of alloHSCT is based on the ability of donor lymphocytes to immunologically control the outgrowth of host residual malignant cells (graft-versus-leukemia [GvL]). However, disease relapse in medullary and/or extramedullary sites remains the most common cause of treatment failure in patients with AML who have had an alloHSCT. The fundamental hypothesis of Project 2 is that events which contribute to relapse and resistance in AML after chemotherapy and allogeneic HSCT are caused by genetic events that can be identified using the discovery platforms established in the Genomics of AML (GAML) PPG. Specific Aim 1: We will define the genetic changes that contribute to AML relapse after chemotherapy. We will perform whole genome, whole exome, and RNA sequencing (Core C) on 50 linked trios of normal skin, cfe novo AML and relapsed AML. Specific Aim 2: We will define the genetic changes that contribute to AML relapse after alloHSCT. Since relapse of AML following alloHSCT can occur at extramedullary (EM) sites with or without bone marrow (BM) involvement, we will investigate 10 patients with BM relapse only, 10 patients with EM relapse only, and 10 patients with both BM and EM relapse. We will purify AML blasts by sorting CD45''?/SSC' AML cells from the de novo sample and post-alloHSCT (BM and/or EM) relapsed samples, and perform whole genome, whole exome, and RNA sequencing to identify all relapse-specific mutations. Specific Aim 3: We will define the genetic changes that contribute to AML relapse after chemotherapy or alloHSCT in a mouse model of AML. We will use a mouse model of acute promyelocytic leukemia (APL) to generate chemotherapy (cytarabine and daunorubicin) and allogeneic T cell resistant tumor cells in vivo . These chemotherapy and allo-resistant APL cells will then be analyzed by whole exome sequencing (Core C), and mRNA and miRNA expression profiling (Core B) to identify genomic and transcriptional alterations that contribute to relapse after chemotherapy vs. alloHSCT.